OPAE C API Programming Guide

Overview

The OPAE C library (libopae-c) is a lightweight user-space library that provides abstraction for FPGA resources in a compute environment. Built on top of the driver stack that supports the FPGA device, the library abstracts away hardware specific and OS specific details and exposes the underlying FPGA resources as a set of features accessible from within software programs running on the host. These features include the acceleration logic preconfigured on the device, as well as functions to manage and reconfigure the device. Hence, the library can enable your applications to transparently and seamlessly take advantage of the FPGA-based acceleration.

Layered

By providing a unified C API, the library supports different kinds of FPGA integration and deployment models, ranging from single-node systems with one or more FPGA devices to large-scale FPGA deployment in a data center. A simple use case, for example, is for a user application running on a system with an FPGA PCIe device to easily use the FPGA to accelerate certain algorithms. At the other end of the spectrum, resource management and orchestration services in a data center can use this API to discover and select FPGA resources and then allocate them to be used by workloads with acceleration needs.

Philosophy

The purpose of OPAE is to provide a common base layer for as wide a range of use cases as possible without sacrificing performance or efficiency. It aims at freeing the developers of applications and frameworks from having to understand the intricacies of the FPGA driver interfaces and FPGA interconnect details by providing a thin abstraction to expose required details of the platform.

To that end, OPAE abstracts access to the key components that frameworks and abstractions need to deal with (for example, FPGA devices and accelerators). It then provides means to interact with these components in the most efficient way possible. Essentially, it tries to provide friendly and consistent interfaces to crucial components of the platform. At the same time, OPAE tries not to constrain frameworks and applications by making optimizations that do not translate to many use cases - and where it does provide convenience functions or optimizations, these are optional.

For example, OPAE provides an interface to allocate physically contiguous buffers in system memory that can be shared between user-space software and an accelerator. This interface enables the most basic feature set of allocating and sharing a large page of memory in one API call; it however does not provide a malloc()-like interface backed by a memory pool or slab allocator. These kinds of optimizations and added functionality are left to higher layers of the software stack, which is better suited to make domain-specific optimizations.

Some Key Concepts

The following key concepts are essential for writing code using the OPAE C API. These concepts are modeled with corresponding data structures and functions in the API specification, as discussed in the Object model section.

  • FPGA: Field Programmable Gate Array is a discrete or integrated peripheral device connecting to a host CPU via PCIe or other type of interconnects.
  • AFU: Accelerator Function Unit, is a computation logic preconfigured on FPGA with the purpose of accelerating certain computation. It represents a resource discoverable and usable by user applications. The logic is designed in RTL and synthesized into a bitstream. A tool (fpgaconf) is provided to reconfigure an FPGA using a bitstream.
  • Accelerator Function (AF): A bitstream for an application-specific accelerator logic, for example, compression, encryption, and mathematical operations.
  • Accelerator: An allocatable accelerator function implemented in an FPGA, closely related to an AFU. An accelerator tracks the ownership of an AFU (or part of it) for a process that uses it. An accelerator can be shared by multiple processes.
  • Shared memory buffers: Memory buffers allocated in user process memory on the host to be shared with an accelerator on the FPGA. Shared memory buffers fascilitate data transfers between the user process and the accelerator it owns.
  • Events: Events are asynchronous notification mechanism. The FPGA driver triggers certian events to indicate error conditions. An accelerator logic can also define its own events. User applications can choose to be notified when certain types of the events occur and respond accordingly.
  • Reconfiguration: An AFU can be replaced by another AFU by a user application that has appropriate privilege.

Use the Sample Code

The library source includes two code samples. Use these samples to learn how to call functions in the library. Build and run these samples as quick sanity checks to determine if your installation and environment are set up properly.

For details about using the sample code, refer to the Running the Hello FPGA Example chapter in the Intel® Acceleration Stack for Intel® Xeon® CPU with FPGAs Getting Started Guide.

High-Level Directory Structure

When successfully built and installed, you can see the following directory structure. This discussion is using installation on Unix/Linux systems as an example. However, it is a similar situation on Windows and MacOS installations.

Directory & Files Contents
include/opae Directory containing all header files
include/opae/fpga.h Top-level header for user code to include
include/opae/access.h Header file for accelerator acquire/release, MMIO, memory management, event handling, etc.
include/opae/bitstream.h Header file for bitstream manipulation functions
include/opae/common.h Header file for error reporting functions
include/opae/enum.h Header file for AFU enumeration functions
include/opae/manage.h Header file for FPGA management functions
include/opae/types.h Various type definitions
lib Directory containing shared library files
lib/libopae-c.so The shared dynamic library for user application to link against
doc Directory containing API documentation
doc/html Directory for documentation of HTML format
doc/latex Directory for documentation of LaTex format
doc/man Directory for documentation of Unix man page format

Basic Application Flow

The picture below depicts the basic application flow from the viewpoint of a user-process. API components are discussed in the next section. The hello_fpga.c sample code is a good example showing the flow in action.

Basic

API Components

The API is designed around an object model that abstracts physical FPGA device and functions available on the device. The object model is not tied to a particular type of FPGA product. Instead, it is a generalized model and can be extended to describe any type of FPGAs.

Object Model

  • fpga_objtype: An enum type to represent the type of an FPGA resource, which is either FPGA_DEVICE or FPGA_ACCELERATOR. An FPGA_DEVICE object is corresponding to a physical FPGA device. Only FPGA_DEVICE objects can invoke management function. FPGA_ACCELERATOR represents an instance of an AFU.
  • fpga_token: An opaque type to represent a resource known to, but not necessarily owned by, the calling process. The calling process must own a resource before it can invoke functions of the resource.
  • fpga_handle: An opaque type to represent a resource owned by the calling process. API functions fpgaOpen() and fpgaClose() (see the Functions section) acquire and release ownership of a resource represented by an fpga_handle.
  • fpga_properties: An opaque type for a properties object. Your applications use these properties to query and search for resources that suit their needs. The properties visible to your applications are documented in the FPGA Resource Properties section.
  • fpga_event_handle: An opaque handle used by the FPGA driver to notify your application about an event, and used by the your application to wait for the notification of the event.
  • fpga_event_type: An enum type to represent kinds of events which can be FPGA_EVENT_INTERRUPT, FPGA_EVENT_ERROR, or FPGA_EVENT_POWER_THERMAL.
  • fpga_result: An enum type to represent the result of an API function. If the function returns successfully the result is FPGA_OK. Otherwise, the result is one of the error codes. Function fpgaErrStr() can translate an error code into human-readable strings.

Functions

The table below groups key API functions by their purposes. For more information about each of the functions, refer to the OPAE C API reference manual.

Purpose Functions Note
Enumeration fpgaEnumerate() Query FPGA resources that match certain properties
Enumeration: Properties fpga[Get|Update|Clear|Clon e|Destroy]Properties]() Manage fpga_prope rties life cycle
  ``fpgaPropertiesGet[Prop]()` ` Get a certain property Prop, from the `FPGA Resource Properties < #fpga-resour ce-propertie s>`__ chapter
  ``fpgaPropertiesSet[Prop]()` ` Set a certain property Prop, from the `FPGA Resource Properties < #fpga-resour ce-propertie s>`__ chapter
Access: Ownership fpga[Open|Close]() Acquire/rele ase ownership
Access: Reset fpgaReset() Reset an accelerator
Access: Event handling fpga[Register|Unregister]E vent() Register/unr egister an event to be notified about
  fpga[Create|Destroy]EventH andle() Manage fpga_event _handle life cycle
Access: UMsg fpgaGetNumUmsg(), fpgaSetUmsgAttributes(), fpgaTriggerUmsg(), fpgaGetUmsgPtr() Low-latency accelerator notification mechanism
Access: MMIO fpgaMapMMIO(), fpgaUnMapMMIO() Map/unmap MMIO space
  fpgaGetMMIOInfo() Get information about a particular MMIO space
  fpgaReadMMIO[32|64]() Read a 32-bit/64-bi t value from MMIO space
  fpgaWriteMMIO[32|64]() Write a 32-bit/64-bi t value to MMIO space
Memory management: Shared memory fpga[Prepare|Release]Buffe r() Manage memory buffer shared between the calling process and an accelerator
  fpgaGetIOVA() Return the virtual address of a shared memory buffer
Management: Reconfiguration fpgaReconfigureSlot() Replace an existing AFU with a new one
Error report fpgaErrStr() Map an error code to a human readable string

FPGA Resource Properties

These are the properties of a resource that can be queried by a user-application, by plugging property name for Prop in the names of fpgaPropertiesGet[Prop]() and fpgaPropertiesSet[Prop]() functions.

Property FPGA accelerat or Note
Parent No Yes ``fpga_token `` of the parent object
ObjectType Yes Yes The type of the resource: either FPGA_DEVIC E or FPGA_ACCEL ERATOR
Bus Yes Yes The bus number
Device Yes Yes The PCI device number
Function Yes Yes The PCI function number
SocketId Yes Yes The socket ID
DeviceId Yes Yes The device ID
NumSlots Yes No Number of AFU slots available on an FPGA_DEVIC E resource
BBSID Yes No The FPGA Interface Manager (FIM) ID of an FPGA_DEVIC E resource
BBSVersion Yes No The FIM version of an FPGA_DEVIC E resource
VendorId Yes No The vendor ID of an FPGA_DEVIC E resource
Model Yes No The model of an FPGA_DEVIC E resource
LocalMemorySize Yes No The local memory size of an FPGA_DEVIC E resource
Capabilities Yes No The capabilities of an FPGA_DEVIC E resource
GUID Yes Yes The GUID of an FPGA_DEVIC E or FPGA_ACCEL ERATOR resource
NumMMIO No Yes The number of MMIO space of an FPGA_ACCEL ERATOR resource
NumInterrupts No Yes The number of interrupts of an FPGA_ACCEL ERATOR resource
AcceleratorState No Yes The state of an FPGA_ACCEL ERATOR resource: either FPGA_ACCEL ERATOR_ASSIG NED or FPGA_ACCEL ERATOR_UNASS IGNED

OPAE C API Return Codes

The OPAE C library returns one of these codes for every public API function exported. Usually, FPGA_OK denotes successful completion of the requested operation, while any return code other than FPGA_OK indicates an error or other deviation from the expected behavior. When using the OPAE C API, always check the API return codes and not the functions that failed output parameters.

Error Code Description
FPGA_OK Operation completed successfully
FPGA_INVALID_PARAM Invalid parameter supplied
FPGA_BUSY Resource is busy
FPGA_EXCEPTION An exception occurred
FPGA_NOT_FOUND A required resource was not found
FPGA_NO_MEMORY Not enough memory to complete operation
FPGA_NOT_SUPPORTED Requested operation is not supported
FPGA_NO_DRIVER Driver is not loaded
FPGA_NO_DAEMON FPGA Daemon (fpgad) is not running
FPGA_NO_ACCESS Insufficient privileges or permissions
FPGA_RECONF_ERROR Error while reconfiguring FPGA

Usage Models

This section illustrates a few typical API usage models with code snippets.

Query and Search for a Resource

The user-code first populates an fpga_properties object with desired properties. Afterwards, fpgaEnumerate() is accessed to search for matching resources.

.. note::

fpgaEnumerate() may return more than one matching resources.

```

#include "fpga/fpga.h"

fpga_guid               guid;
fpga_properties         filter = NULL;
fpga_result             res;
fpga_token              tokens[MAX_NUM_TOKENS];
uint32_t                num_matches = 0;

/* Start with an empty properties object */
res = fpgaGetProperties(NULL, &filter);

/* Populate the properties object with desired values.
   In this case, we want to search for accelerators that match a
   particular GUID.
*/
uuid_parse(GUID, guid);
res = fpgaPropertiesSetObjectType(filter, FPGA_ACCELERATOR);
res = fpgaPropertiesSetGuid(filter, guid);

/* Query the number of matched resources */
res = fpgaEnumerate(&filter, 1, NULL, 1, &num_matches);

/* Return all matched resources in tokens */
res = fpgaEnumerate(&filter, 1, tokens, num_matches, &num_matches);

/* Destroy the properties object */
res = fpgaDestroyProperties(&filter);

/* More code */
......

/* Destroy tokens */
for (uint32_t i = 0; i < num_matches; ++i) {
    res = fpgaDestroyToken(tokens[i]);
}
The `fpgaEnumerate()` function can take multiple `fpga_properties`
objects (in an array). In this situation, the function returns resources that
match *any* of the properties object. In other words, the multiple properties
objects are logically OR'ed in the query operation.
* Again, `fpga_token` objects return by `fpgaEnumerate()` do *not* signify
ownership. To acquire ownership of a resource represented by a token, pass the
token to `fpgaOpen()`.

Acquire and Release a Resource

Acquiring and releasing ownership of a resource is done using fpgaOpen() and fpgaClose(). The calling process must own the resource before it can do MMIO, share memory buffers, and use functions offered by the resource.

#include "fpga/fpga.h"

fpga_handle             handle;
fpga_result             res;

/* Acquire ownership of a resource that was previously returned by
   `fpgaEnumerate()` as a token
*/
res = fpgaOpen(token, &handle);

/* More code */
......

/* Release the ownership */
res = fpgaClose(handle);

Shared Memory Buffer

This code snippet shows how to prepare a memory buffer for sharing between the calling process and an accelerator.

#include "fpga/fpga.h"

fpga_handle             handle;
fpga_result             res;

/* Hint for the virtual address of the buffer */
volatile uint64_t       *addr_hint;
/* An ID we can use to reference the buffer later */
uint32_t                bufid;
/* Flag to indicate if the buffer is preallocated or not */
int                     flag = 0;

/* Allocate (if necessary), pin, and map a buffer to be accessible
   by an accelerator
*/
res = fpgaPrepareBuffer(handle, BUF_SIZE, (void **) &addr_hint,
                        &bufid, flag);

/* The actual address mapped to the buffer */
uint64_t                iova;
/* Get the IO virtual address for the buffer */
res = fpgaGetIOVA(handle, bufid, &iova);

/* Inform the accelerator about the virtual address by writing to its mapped
   register file
*/
......

/* More code */
......

/* Release the shared buffer */
res = fpgaReleaseBuffer(handle, bufid);
The `flag` variable can take a constant `FPGA_BUF_PREALLOCATED`, which
indicates that the address space pointed to by `addr_hint` is already allocated
by the calling process.

MMIO

This code snippet shows how to map/unmap the register file of an accelerator into the virtual memory space of the calling process.

#include "fpga/fpga.h"

fpga_handle             handle;
fpga_result             res;

/* Index of the MMIO space. There might be multiple spaces on an accelerator */
uint32_t                mmio_num = 0;
/* Mapped address */
uint64_t                mmio_addr;

/* Map MMIO */
res = fpgaMapMMIO(handle, mmio_num, &mmio_addr);

/* Write to a 32-bit value to the mapped register file at a certain byte
   offset.

   CSR_CTL is the offset in the mapped space to where the value will be
   written. It's defined elsewhere.
*/
res = fpgaWriteMMIO32(handle, mmio_num, CSR_CTL, value);

/* More code */
......

/* Unmap MMIO */
res = fpgaUnmapMMIO(handle, mmio_num);
Every AFU has its own layout of register spaces and its own protocol about
how to control its behavior through the registers. These are defined in the
AF used to implemented the AFU.